Response to: ‘Association between migraine and cardiovascular mortality: Is there a temporal trend?’ by Rossi et al.

Migraine with aura has been associated with cardiovascular disease (CVD), in particular with ischemic stroke (1); however, such an association has not been consistently shown for CVD mortality (2).

Rossi et al. (3) have reinvestigated data from observational studies on the association between migraine and mortality as summarized in our meta-analysis (2). Specifically, they have addressed the question of whether there is a time trend for the risk of CVD mortality by performing a meta-regression with the relative risk as the dependent variable and the publication date as the independent variable. They report a suggestive trend towards increased CVD mortality over time (p = 0.054).

In general, when assessing time trends, the first step involves identifying the observation of interest (here mortality) by an appropriate time period. This has three important implications. First, the concept of an appropriate time period means that the population should be under observation during the time period under question. Second, the time period of observation should be of reasonable duration. Although very short time windows may be suitable for post-operative outcomes (4,5), for other outcomes this may not be appropriate, as the number of events may be too few to calculate a reliable risk estimate. For studies investigating the association between migraine and mortality long-term follow-up is needed. Finally, the observation periods among the studies should be of similar length in order to be comparable.

With respect to the studies investigating the association between migraine and mortality (2), the following aspects need to be considered. Rossi et al. (3) have chosen the date of publication as a proxy for the observation period. However, the time periods from end of observation until publication vary between 2 and 5 years. Further, the observation periods among the studies vary considerably ranging from an average of 1.4 to 25.9 years, with three of the studies spanning observation times ≥10 years (2). Hence, time trends within these observation periods are undetectable when considering summary risk estimates. Further, slicing long observation periods into periods of equal length may improve comparability; however, this may interfere with the ability to accurately assess the risk estimate.

Given these considerations we have not attempted to investigate a time trend in our analysis (2), and we do not believe that such an analysis validly reflects any potential change in risk over time. We agree, however, that future studies are needed to clarify the association of migraine and migraine specifics with overall and cause-specific mortality.

Complete disclosures for the last 2 years

Markus Schürks has received an investigator-initiated research grant from the Migraine Research Foundation and a honorarium from L.E.K. Consulting for telephone surveys. Since August 2011 he is a full-time employee of Bayer HealthCare Germany.

Pamela Rist has received funding from a training grant from the US National Institutes of Health and from the Rose Traveling Fellowship Program in Chronic Disease Epidemiology and Biostatistics from the Harvard School of Public Health.

Tobias Kurth has received investigator-initiated research funding from the French National Research Agency, the US National Institutes of Health, the Migraine Research Foundation, and the Parkinson’s Disease Foundation. He is a consultant to World Health Information Science Consultants, LLC, and has received honoraria from Allergan, the American Academy of Neurology, and Merck for educational lectures and from MAP Pharmaceutical for contributing to a scientific advisory panel.